Lykke Business. Algo Store. ‹ Prime Holding JSC

Lykke Business. Algo Store.

Automated Trading at Speed

Overview

Lykke is an international, Swiss-based company building a global marketplace for the free exchange of financial assets. They wanted to build a platform that allows users to develop their automated trading strategies, creating unique and practical algorithms and using most crypto or forex assets on the market. They ought to be able to safely test and execute their algorithms on historical data without any risk of financial loss, evaluating the potential effectiveness of their strategy before deploying it live with real assets.

Challenges

AlgoStore needed to offer a competitive automated trading software solution on the market. To achieve it, we had to build a system that provides faster transactions and higher customization whilst having an easy-to-use interface suitable for users with a wide variety of skills and experience. Since Lykke was growing quickly and their customers' expectations were constantly evolving, the team needed to be very flexible and resourceful to adapt rapidly to changing requirements and build a stable and reliable system simultaneously.

Solution

We built a complex system providing users with a simplified interface for creating and managing trading algorithms. We used a cloud-based distributed microservices architecture with the latest technologies - .NET Core 2.1 for the APIs and a modern front-end framework Angular 4/5 for the user interface to achieve speed, reliability, and convenience of use.

To achieve scalability at a lower cost, we used Microsoft Azure Storage as the default storage. We implemented a RESTful API for greater flexibility and enabled High-Frequency Trading by utilizing TCP sockets for faster service communication. Also, by deploying the whole AlgoStore product in the same infrastructure as the Lykke platform, we achieved a high-speed and highly reliable automated trading platform that is easy to use, maintain and update.

Furthermore, the team built and maintained a robust framework for automation testing and developed many automated tests to ensure stability, security, and quality. To establish easier testing and deployment of pre-production and development environments, we used TeamCity for the CI (continuous integration) and Kubernetes for container orchestration.

Results

An easy-to-use intuitive algo trading interface
Ability to use a favourite programming language such as C#, Java, and Python to create the best and most effective trading strategy
Architecture that guarantees minimal reaction times and supreme performance due to the physical proximity of the servers and the use of the fastest communication protocol

Quick Facts

Duration: 1 year

Technology Stack: .Net Core 2.0 and 2.1, Angular, Microsoft Azure Storage, Kubernetes, TeamCity, Git, RabbitMQ, Redis

Team: Varied over time. 4 - 6 .NET Developers, 1 QA, 1-2 FrontEnd Developers, 1 Architect as a consultant and 1 PM/PO

Let's talk about your Ideas.

Success Stories